Human Augmentation of UAV Cyber-Attack Detection

Operator Strategy Model Development in UAV Hacking Detection

“…Operators were able to successfully detect hacking events such that 65% of total experimental sessions exhibited at least 80% correct hacking identification. We also discovered that operators with significant video game experience were the best performers in hacking detection [9].…”

DeepSIM: GPS Spoofing Detection on UAVs using Satellite Imagery Matching

“…Our proposed approach falls into this category. Methods with similar auxiliary equipment include [10,33,47,68], which will be discussed and compared with our method in Section 7 (Related Work).…”

“…Recently, with the remarkable progress of Computer Vision and Artificial Intelligence, some researchers have considered using vision-based techniques as assistance to detect GPS spoofing attacks. Such out-of-band methods have already been proved effective and successful [10,33,47,68].…”

“…Zhu et al proposed a collaborative approach for a single operator of multiple-UAV supervisory control systems using humanautonomy, in which human geolocation is used to help detect possible UAV cyber-attacks [68]. This method compares the nontempered UAV camera video feed to the potentially falsified GPS location, and detect inconsistencies between the UAV video feed and UAV GPS location.…”

“…The above proposals made innovative use of airborne equipment and resources of UAVs, e. g., cameras, maps, IMUs to design promising out-of-band ways to detect GPS spoofing attacks. For instance, the method proposed by Zhu et al [68], however, relied on human detection and could not be fully automated. Others leveraged traditional hand-crafted visual features with performance impairment [33], or applied visual features to rectify data and compensate the drift from other sensors [10], requiring extra devices (i. e., IMUs) to cooperate with a camera.…”